it. The identification of states as clusters relies on standard clustering algorithms that ignore conformational energy and instead use similarity over low-dimensional representations (a process referred to as featurization in current MSM analysis) of conformations. In this work, we propose to leverage the fact that the conformations sampled in simulation populate an unknown energy landscape. The neighborhood structure of the landscape can be extracted via spatial statistical analysis that does not ignore energy but instead uses it to identify basins and basin-separating saddles in the landscape. We utilize this information to identify states and show that such a definition of states is more robust than that obtained via clustering. We proceed to analyze in this manner various MD trajectories and summarize the conformational dynamics in a comparative manner, showing that a landscape-based identification of conformational states is promising for detecting and summarizing conformational dynamics. DNA-Sequencing of tumor cells has revealed thousands of genetic mutations but cancer is caused by some of them. The challenge is to identify the mutations that contribute to tumor growth which is currently being done manually. A clinical pathologist manually reviews and classifies each genetic mutation based on evidence from clinical literature. This process is very cumbersome and expensive in terms of money and time. We have used different Machine Learning (ML) classifiers (Random Forest, k-Nearest Neighbors, Naive Bayes, Support Vector Machine and Gradient Boosting etc) for multiclass classification of cancer mutations. We used OncoKB dataset which has three features (Gene, Mutation/Variation, Clinical literature). Gene and Variation were encoded to numerical features from categorical features using one-hot encoding. Clinical literature was converted to numerical vectors of fixed length using tf-idf, word2vec and doc2vec and then fed into ML models. We achieved 0.9 logloss and 68% accuracy by XGBClassifier, while the random model gives 2.68 logloss and 12% accuracy. Our model performs significantly better than random model (3x on logloss and 5.67 times on accuracy). This work will assist the pathologist in classifying the genetic mutation in less time and save the patient's life and time by reducing the misdiagnosis rate. Moreover accurate treatment will reduce toxicity in cancer patients. Cellular morphodynamics is a phenotypic outcome of numerous cellular processes, including migration, proliferation, and differentiation. It is commonly utilized as an indicator of physiological and pathological state of cells. Therefore, quantitative characterization on morphodynamic feature of cells is essential for evaluating cell state and understanding the effect of underlying molecular pathway that governing the morphogenetic process. By employing Hilbert-Huang transform (HHT) based spectral decomposition method, we decomposed the convoluted cell edge movement time series into several analyzable functions with instantaneous freque...
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.